Cell Culture and Electroporation of 3T3L1 Adipocytes

The 3T3L1 fibroblasts were grown in DMEM with 10% fetal bovine serum (FBS) and differentiated into adipocytes as described (Baumann et al., 2000). The 3T3L1 adipocytes were transfected with stealth RNA interference (RNAi) duplexes (Invitrogen, Carlsbad, CA) by electroporation. In brief, adipocytes at day 2 of differentiation were detached from culture dishes with 0.25% trypsin, washed twice, and resuspended in phosphate-buffered saline (PBS). Approximately 5 × 10⁶ cells (half of the cells from one p150 dish) were mixed with RNAi duplexes, which were delivered to the cells by electroporation with a Bio-Rad gene pulser II system (Richmond, CA; 0.16 kV and 960 μF). After electroporation, cells were mixed with medium for 10 min before replating.

Immunoprecipitation and Immunoblotting

Cells were lysed with HNTG buffer using 1% Triton X-100. Lysates were incubated with the indicated antibodies for 2 h at 4°C. The immune-complexes were precipitated with protein A or G agarose beads (Amersham Pharmacia, Piscataway, NJ) and washed with lysis buffer. The samples were resolved in 4-20% gradient SDS-PAGE and analyzed by immunoblotting. All immunoblots were developed by enhanced chemiluminesence (Amersham Pharmacia).

Immunofluorescence Microscopy

For immunofluorescence studies, cells were processed as described (Kimura et al., 2001, 2002; Liu et al., 2005). To make plasma membrane sheets, cells were treated for 5 min with the Triton extraction buffer (50 mM MES, pH 6.0, 150 mM NaCl, 0.25% Triton X-100, 1 mM CaCl₂, 0.5 mM MgCl₂) before fixation. To detect myc-Exo70, cells were stained with anti-myc mAb or polyclonal antibody (Santa Cruz Biotechnology). To detect HA-TC10, cells were stained with anti-HA mAb (Santa Cruz Biotechnology). After incubation with primary antibodies, cells were incubated with Alexa⁴⁸⁸ or Alexa⁵⁹⁴ goat anti-mouse or anti-rabbit immunoglobulin (IgG), obtained from Molecular Probes (Invitrogen, Eugene, OR). Images were captured by using an Olympus FV300 confocal laser scanning microscope (Melville, NY).

Expression Constructs

To generate a mammalian expression vector, full-length Sec8 was subcloned into the pKH3 vector. YFP-mycSAP97 construct was a gift from Dr. Benjamin Margolis (University of Michigan). The His-tagged proteins Sec8 and its Δ4aa deletion mutant were cloned into a pET15b vector.

2-Deoxyglucose Uptake Assay

The RNAi-transfected cells were reseeded on 12-well plates and cultured for 4 d. After incubation with DMEM containing 0.5% FBS for 3 h, cell were washed with Krebs-Ringer buffer (130 mM NaCl, 5 mM KCl, 1.3 mM CaCl_2, 1.3 mM MgSO_4, 25 mM HEPES, pH 7.4) and incubated with or without insulin for 30 min. Glucose uptake was initiated by addition of [¹⁴C]2-deoxy-d-glucose to a final assay concentration of 100 μM for 5 min and terminated by three washes with ice-cold PBS, the cells were solubilized with 0.1% SDS, and [¹⁴C] was determined by scintillation counting.

Sucrose Gradient Centrifugation

3T3L1 adipocytes were homogenized in HES buffer (20 mM HEPES, pH 7.5, 250 mM sucrose, 1 mM EDTA) with a Dounce homogenizer. After homogenization, cells were centrifuged at 800 × g and the supernatant was centrifuged at 15,000 × g to produce a total plasma membrane fraction. Total plasma membrane fraction was solubilized in 0.3 ml of 0.2% Triton X-100 in MBS, and made up to 0.4 ml with MBS. These samples were mixed with 0.4 ml of 80% sucrose (wt/vol), and overlaid with 3.5 ml of 30%, 1 ml of 5% sucrose. The gradient was centrifuged at 150,000 × g for 19 h, and 440-μl fractions were collected from the top of each gradient.

Insulin Controls the Localization of the Exocyst Complex

To evaluate the role of lipid rafts in coordinating the assembly of the exocyst complex, we performed a number of immunohistochemistry and fractionation experiments to assay the localization of exocyst proteins in these microdomains. Differentiated 3T3L1 adipocytes were transfected with cDNAs expressing HA-TC10 and its mutants, along with myc-Exo70 and caveolin-eGFP as a marker of lipid rafts. The cells were then stimulated with insulin, and plasma membrane sheets were prepared by a triton extraction method that leaves only lipid raft proteins on the cover glass (Watson et al., 2001). As previously reported (Watson et al., 2001; Chiang et al., 2002), both caveolin and TC10 constitutively reside in detergent-extracted sheets derived from untreated and insulin-stimulated cells. Exo70 protein was not detected in these sheets in untreated cells, but was found in sheets from cells treated with insulin. Moreover, coexpression with constitutively active TC10 (Q⁶⁷L) produced the localization of Exo70 at lipid rafts even without insulin stimulation (Figure 2A). Although we are unable to detect endogenous Sec8, we transfected cells with Myc- or HA-tagged Sec8 and treated cells with or without insulin. Insulin stimulated the translocation of Sec8 to the plasma membrane. This effect of the hormone was mimicked by cotransfection of cells with constitutively active TC10 (Supplementary Figure 1).

Open in a separate window

Figure 2.

Insulin stimulates the translocation of exocyst proteins to lipid rafts. (A) HA-TC10 wild type or TC10 (Q⁶⁷L) was transfected into 3T3L1 adipocytes with myc-Exo70 and caveolin-eGFP. To make plasma membrane sheets, cells were treated for 5 min with the Triton extraction buffer. The plasma membrane sheets were fixed and stained with an anti-HA mAb and anti-myc polyclonal antibody followed by Alexa⁶³³ goat anti-mouse IgG and Alexa⁵⁶⁸ goat anti-rabbit IgG. The cells transfected with HA-TC10WT were starved overnight and treated with or without insulin for 5 min (left 8 panels). The cells transfected with HA-TC10 Q⁶⁷L were not starved or stimulated by insulin (right 4 panels). (B) After a 5-min insulin stimulation, 3T3L1 adipocytes were harvested in the nondetergent buffer followed by sucrose density gradient fractionation. The expression of caveolin, transferrin receptor and Sec8 were evaluated by immunoblot. (C) Cells treated with insulin for the indicated times were lysed and fractionated on sucrose gradients. Fractions 4 and 5 were pooled as the lipid raft fraction, and fraction 10-12 were pooled as the nonlipid raft fraction. These samples were resolved in 4-20% gradient SDS-PAGE and analyzed by immunoblotting with anti-Exo70, Sec8, and caveolin antibodies.

To confirm the localization of exocyst proteins in lipid rafts, we performed sucrose density gradient fractionation of purified plasma membranes prepared from untreated and insulin-treated cells, followed by immunoblot analysis, as previously described (Chamberlain and Gould, 2002; Chiang et al., 2002; Kimura et al., 2002). Adipocytes were treated with or without insulin for 5 min, followed by lysis and fractionation. Caveolin1 was exclusively detected in the lipid raft-enriched fraction 4 and 5 of the gradient and was unaffected by insulin treatment, as previously shown (Chamberlain and Gould, 2002). In contrast, the transferrin receptor was exclusively found at the bottom of the gradient in fractions 10-12, where it increased in response to insulin (Chamberlain and Gould, 2002). Interestingly, Sec8 was detected in both raft (fractions 4 and 5) and nonraft fractions (10-12) in untreated cells. However, both proteins were increased only in the lipid raft fractions in cells treated with insulin (Figure 2B).

To analyze the kinetics of this effect of insulin in more detail, we performed a time course of insulin stimulation. After hormone treatment, the plasma membranes were purified and fractionated further on sucrose gradients, and the raft and nonraft fractions were pooled before immunoblotting. Although the proteins residing in the nonraft fractions did not change as a consequence of insulin stimulation, treatment of cells with insulin produced the rapid and transient translocation of both Exo70 and Sec8 into the lipid raft fraction. This effect of insulin was maximal at 5 min and declined thereafter (Figure 2C).

Glut4 Vesicles Transit through Lipid Rafts before Fusion

Glut4-containing vesicles translocate to the plasma membrane in response to insulin stimulation (Saltiel and Kahn, 2001). However, whether these vesicles are targeted to discrete sites on the plasma membrane is not known. Because the exocyst appears to assemble in large part at lipid rafts, we sought to evaluate whether Glut4 might be targeted to these microdomains early in the translocation process. Thus, to evaluate the kinetics of Glut4 translocation to the plasma membrane, we examined the time course of Glut4 translocation using the fractionation method described above and determined the concentration of Glut4 in lipid raft and nonlipid raft fractions of the plasma membrane after insulin stimulation (Figure 3A). Glut4 was rapidly translocated to rafts in response to insulin. This effect was observed in as little as 2 min and reached maximal levels 10 min after insulin stimulation. The translocation of Glut4 into raft fractions preceded the appearance of the protein in nonlipid raft fractions, where peak levels were detected 30 min after insulin stimulation. Thus, these data suggest two possible models: one in which Glut4 is initially translocated to lipid rafts, and then transits to nonraft fractions, perhaps after fusion of the vesicles, and a second in which there are two pools of Glut4, a fast pool targeted to rafts and a slower pool target to nonraft regions.

Open in a separate window

Figure 3.

Time course of Glut4 translocation to lipid rafts. (A) Glut4 expression in lipid raft, nonlipid raft, or nonplasma membrane fractions were examined at different time points after insulin stimulation. The intensity of Glut4 levels in the raft, nonraft, and nonplasma membrane were quantified. (B) The lipid raft and nonlipid raft fractions purified by sucrose gradient from the different conditions (with or without insulin stimulation, Exo70 RNAi or scrambled oligo-transfected as indicated) were loaded on the same gel and blotted with anti-Exo70, Sec8, Glut4, transferrin receptor, and caveolin antibodies. (C) Four days after transfection of Exo70 or scrambled RNAi oligo, cells were treated with insulin for 10 min and incubated with Triton extraction buffer to make plasma membrane sheets. The plasma membrane sheets were fixed and stained with anti-Glut4 polyclonal antibody and anti-caveolin2 mAb followed by Alexa⁵⁹⁴ goat anti-rabbit IgG and Alexa⁴⁸⁸ goat anti-mouse IgG. Arrows show the rosette structures.

The Exocyst Complex Directs Glut4-containing Vesicles to Lipid Rafts

Because both Glut4 and exocyst proteins translocate to lipid rafts in response to insulin, we wondered whether the assembly of the exocyst complex plays a crucial role in the trafficking of Glut4 vesicles to these microdomains in the adipocyte. We therefore used RNAi to selectively knockdown exocyst proteins and evaluated their targeting to lipid raft fractions. Exo70 knockdown reduced the insulin-stimulated translocation of both Glut4 and Sec8 into the lipid raft fraction, but did not markedly affect the levels of either protein in the nonraft fractions, nor did it affect the translocation of the transferrin receptor (Figure 3B).

We next examined the effect of Exo70 knockdown on the translocation of Glut4 to detergent-extracted plasma membrane sheets. Cells were electroporated with Exo70 RNAi or a scrambled oligo and then treated with or without insulin for 10 min. In scrambled oligo-transfected cells, endogenous Glut4 and caveolin costained in circular rosette structures in plasma membrane sheets, suggesting that a subpopulation of Glut4 vesicles translocate to lipid rafts after insulin stimulation (Figure 3C). Knockdown of Exo70 disrupted the colocalization of Glut4 and caveolin in rosettes, suggesting that the exocyst complex selectively tethers Glut4 to lipid rafts in the early phase of insulin signaling.

We previously reported that an Exo70 mutant containing only the N-terminal sequences (Exo70-N) inhibited insulin-stimulated glucose uptake (Inoue et al., 2003). To evaluate the mechanism of action of this mutant form of Exo70, we examined its localization in the plasma membrane. 3T3L1 adipocytes were electroporated with cDNAs encoding myc-Exo70-WT and myc-Exo70-N or a vector control and treated with or without insulin for 5 min. Subcellular localization of transfected and endogenous exocyst proteins was then evaluated by sucrose density gradient fractionation (Figure 4A). As shown above in untransfected cells, levels of Exo70, Sec8, and Glut4 were low in lipid raft fractions from unstimulated cells expressing Exo70-WT, but increased upon insulin stimulation. In contrast, the overexpression of myc-Exo70-N in adipocytes suppressed the insulin-dependent translocation of Exo70, Sec8, and Glut4 to lipid raft fractions. These data suggest that endogenous Exo70 and other components of the exocyst complex are competitively displaced from lipid rafts by overexpression of the dominant-negative form of the protein.

Open in a separate window

Figure 4.

Dominant negative Exo70 inhibits Glut4 movement to lipid rafts. (A) Myc-Exo70-WT and -N were overexpressed in 3T3L1 adipocytes by electroporation and harvested 3 d later without starvation or stimulation by insulin. Sucrose gradients fractionations were performed as described above, and fractions were blotted with anti-myc mAb and anti-caveolin antibody. (B) Vector control, myc-Exo70-WT and -N overexpressed 3T3L1 adipocytes were treated with or without insulin for 5 min, and sucrose gradient fractionation were performed as in A. The fractions were loaded in SDS-PAGE and blotted with anti-Sec8 mAb, anti-Exo70 mAb, anti-Glut4 polyclonal antibody, and anti-caveolin polyclonal antibody.

We thank Dr. Shu-Chan Hsu (Rutgers University) for the anti-Exo70 mAb, Dr. Benjamin Margolis (University of Michigan) for the YFP-SAP97 construct, and Zhao-Qin Bao for technical assistance. This work was supported by Grants DK61618 and DK60591 from the National Institutes of Health.